quelques applications des fonctions `a variation born´ee
TRANSCRIPT
![Page 1: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/1.jpg)
Quelques applications des fonctions a variationbornee en dimension finie et infinie.
These de Doctorat
Michael Goldman
CMAP, Polytechnique
9 decembre 2011
![Page 2: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/2.jpg)
Topic of the Thesis
![Page 3: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/3.jpg)
Introduction
Primal-Dual methods in image processing
Sets with prescribed mean curvature in periodic media
Variational problems in Wiener spacesIntroduction to Wiener spacesApproximation of the perimeter in Wiener spacesConvexity of minimizers of variational problems in Wiener spaces
![Page 4: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/4.jpg)
Introduction
![Page 5: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/5.jpg)
Introduction
Functions of bounded variation have a central position in manyproblems in the Calculus of Variations.
DefinitionLet u ∈ L1(Ω) then u ∈ BV (Ω) if
∫
Ω|Du| := sup
ϕ∈C1c (Ω)
|ϕ|∞≤1
∫
Ωu divϕ < +∞.
DefinitionA set E ⊂ R
m is called a set of finite perimeter if
P(E ) :=
∫
Rm
|DχE | < +∞.
If E is a smooth set then P(E ) = Hm−1(∂E ).
![Page 6: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/6.jpg)
Typical functions in BV
−2 −1.5 −1 −0.5 0 0.5 1 1.5 2−2
0
2
−0.5
0
0.5
![Page 7: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/7.jpg)
Primal-Dual methods in image
processing
![Page 8: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/8.jpg)
Examples of problems in image processing
Inpainting
Deblurring
![Page 9: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/9.jpg)
Many problems in image processing can be model as solving aminimization problem :
J(u) :=
∫
Ω|Du|+ G (u)
where G is a lsc convex function on L2.
Example : denoising with ROF corresponds to G (u) = λ2
∫Ω |u− f |2
It can also be used for inpainting, deblurring, zooming...
![Page 10: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/10.jpg)
Many problems in image processing can be model as solving aminimization problem :
J(u) :=
∫
Ω|Du|+ G (u)
where G is a lsc convex function on L2.
Example : denoising with ROF corresponds to G (u) = λ2
∫Ω |u− f |2
It can also be used for inpainting, deblurring, zooming...
Problem : How to solve the minimization problem ?
![Page 11: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/11.jpg)
Idea of the method
Remind : The total variation is defined as
∫
Ω|Du| = sup
ξ∈C1c (Ω)
|ξ|∞≤1
−
∫
Ωu div ξ
Hence the minimization problem rewrites
minu∈BV
J(u) = minu∈BV
supξ∈C1c (Ω)
|ξ|∞≤1
−
∫
Ωu div ξ + G (u)
⇒ It is thus equivalent to finding a saddle point
![Page 12: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/12.jpg)
The Arrow-Hurwicz method for finding saddle points
For a function K , this method is
∂u∂t
= −∇uK (u, ξ)
∂ξ∂t
= ∇ξK (u, ξ)
It is a gradient descent in the primal variable u and a gradientascent in the dual variable ξ.
![Page 13: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/13.jpg)
When K (u, ξ) = −
∫
Ωu div ξ + G (u) we find
∇uK = − div ξ + ∂G (u)
∇ξK = Du
![Page 14: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/14.jpg)
When K (u, ξ) = −
∫
Ωu div ξ + G (u) we find
∇uK = − div ξ + ∂G (u)
∇ξK = Du
which formally amounts to solve :
∂u∂t
= div ξ − ∂G (u)
∂ξ∂t
= Du |ξ|∞ ≤ 1
This method proposed by Appleton and Talbot is the continuousanalogous of the method proposed by Chan and Zhu in thediscrete setting.
![Page 15: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/15.jpg)
TheoremGiving appropriate meaning to the previous system, there exists a
unique solution to the Cauchy problem.
Moreover, for G (u) = λ2 |u − f |2
L2there is convergence towards the
minimizer u of J and we have the a posteriori estimation
|u − u| ≤1
2
|∂tu|
λ+
√|∂tu|2
λ2+
8|Ω|12
λ|∂tξ|
![Page 16: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/16.jpg)
This extends to :
segmentation with geodesic active contours,
J(u) =
∫
Ωg(x)|Du|
problems with boundary conditions
![Page 17: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/17.jpg)
Interest of this continuous approach
Give a better understanding of the discrete method
Lead to new results also in the discrete setting (such as aposteriori estimates)
Give rise to more isotropic results (absence of discretizationbias)
![Page 18: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/18.jpg)
Restauration by AT left and CZ right
![Page 19: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/19.jpg)
Zoom on the top right corner
![Page 20: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/20.jpg)
Numerical results
![Page 21: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/21.jpg)
Sets with prescribed mean
curvature in periodic media
![Page 22: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/22.jpg)
The Problem
Let g : Rm → R periodic, find acompact set with
κ = g
![Page 23: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/23.jpg)
The Problem
Let g : Rm → R periodic, find acompact set with
κ = g
Example : If g ≡ C > 0 then a solution is given by a ball.
![Page 24: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/24.jpg)
Main result
In general there is no solution but
TheoremLet g be periodic with zero mean and sufficiently small norm then
for every ε there exists ε′ ∈ [0, ε] and a compact solution to
κ = g + ε′.
![Page 25: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/25.jpg)
Idea of proof
Consider the volume constrained problem
f (v) := min|E |=v
P(E )−
∫
E
g
then
Proposition
For every v > 0 there exists a compact minimizer Ev of this
problem.
![Page 26: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/26.jpg)
Ev satisfies the Euler-Lagrange equation
κ = g + λ
we can prove f is Lipschitz and
f ′(v) = λ
![Page 27: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/27.jpg)
Ev satisfies the Euler-Lagrange equation
κ = g + λ
we can prove f is Lipschitz and
f ′(v) = λ
⇒ if we can find v with 0 ≤ f ′(v) < ε we are done.
![Page 28: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/28.jpg)
Ev satisfies the Euler-Lagrange equation
κ = g + λ
we can prove f is Lipschitz and
f ′(v) = λ
⇒ if we can find v with 0 ≤ f ′(v) < ε we are done.
Since f ≈ cvm−1m , we can find vn → +∞ with f ′(vn) → 0.
![Page 29: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/29.jpg)
Behaviour of large volume minimizersThere exists φg a one-homogeneous convex function such thatletting
Wg =
x ∈ R
m : maxφg (y)≤1
x · y ≤ 1
φg = | · |∞ φg = | · |2 φg = | · |1
Theorem
Ev =(|Wg |v
) 1mEv + zv
it holds limv→+∞
∣∣∣Ev∆Wg
∣∣∣ = 0.
![Page 30: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/30.jpg)
Variational problems in Wiener
spaces
![Page 31: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/31.jpg)
The Wiener space
DefinitionA Wiener space is a Banach space X with a Gaussian measure γ.
if x∗ ∈ X ∗ then x → 〈x∗, x〉 ∈ L2γ(X ).
H := X ∗L2γ(X )
.
![Page 32: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/32.jpg)
Canonical cylindrical approximation
Let (x∗i )i∈N ∈ X ∗ be an orthonormal basis of H. Then
Πm(x) := (〈x∗1 , x〉, . . . , 〈x∗m, x〉).
Gives a decomposition of X ∼= Rm ⊕ X⊥
m and γ = γm ⊗ γ⊥m , withγm, γ
⊥m Gaussian measures on R
m, X⊥m respectively.
For u ∈ L1γ(X ),
Emu(x) :=
∫
X⊥m
u(Πm(x), y)dγ⊥m(y).
![Page 33: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/33.jpg)
Gradient and Divergence
In this context one can define a notion of gradient and divergencesuch that
Proposition
For smooth enough Φ and u,
∫
X
u divγ Φ dγ = −
∫
X
[∇u,Φ] dγ.
⇒ This gives definitions of Sobolev and BV functions as in theEuclidiean setting.We say that E is of finite Gaussian perimeter if
Pγ(E ) :=
∫
X
|DγχE | < +∞.
![Page 34: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/34.jpg)
Isoperimetric inequality
Definition
Φ(t) = γ(〈x∗m, x〉 ≤ t)
α(v) = Φ−1(v).
U(v) = Pγ(〈x∗m, x〉 ≤ α(v)).
0
α(v)
xm
E = 〈x∗m, x〉 ≤ α(v)
Theorem (Isoperimetric inequality)
The half spaces are the only isoperimetric sets i.e. for every set E ,
Pγ(E ) ≥ U(γ(E )).
![Page 35: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/35.jpg)
The Ehrhard symetrization
DefinitionLet E ⊂ X and m ∈ N. The Ehrhard symmetral of E is :
xm
Em−1χE (x)
E
E s
α(Em−1χE )
x
Em−1χE (x)
![Page 36: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/36.jpg)
The Ehrhard symetrization
DefinitionLet E ⊂ X and m ∈ N. The Ehrhard symmetral of E is :
xm
Em−1χE (x)
E
E s
α(Em−1χE )
x
Em−1χE (x)
Pγ(Es) ≤ Pγ(E )
![Page 37: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/37.jpg)
Approximation of the perimeter
in Wiener spaces
![Page 38: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/38.jpg)
A Modica-Mortola result
Idea : approximate the perimeter with
Jε(u) :=
∫
X
ε
2|∇u|2 +
W (u)
εdγ
where W is a double-well potential.
![Page 39: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/39.jpg)
A Modica-Mortola result
Idea : approximate the perimeter with
Jε(u) :=
∫
X
ε
2|∇u|2 +
W (u)
εdγ
where W is a double-well potential.
Problem : No compactness inthe strong L2γ(X ) topology
⇒ we have to consider the weak topology.
![Page 40: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/40.jpg)
A Modica-Mortola result
Idea : approximate the perimeter with
Jε(u) :=
∫
X
ε
2|∇u|2 +
W (u)
εdγ
where W is a double-well potential.
Problem : No compactness inthe strong L2γ(X ) topology
⇒ we have to consider the weak topology.
But the perimeter is NOT lsc for this topology...
![Page 41: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/41.jpg)
We thus have to compute the relaxation of the perimeter :
F (u) := inf limPγ(En) / En u.
![Page 42: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/42.jpg)
We thus have to compute the relaxation of the perimeter :
F (u) := inf limPγ(En) / En u.
Since Pγ(Es) ≤ Pγ(E ),
F (u) = inf limPγ(Esn ) / E s
n u.
![Page 43: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/43.jpg)
Relaxation of the perimeter
If u(x) = u(〈x∗1 , x〉) and Em = 〈x∗m, x〉 ≤ α u(〈x∗1 , x〉) :
xm
α(u(x1))
Em
x1
Em u and Pγ(E ) =
∫
R
√U2(u) + |Dγu|2dγ1.
![Page 44: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/44.jpg)
Main results
TheoremThe relaxation of the perimeter for the weak L2γ(X ) topology is
given by
F (u) =
∫
X
√U2(u) + |Dγu|2 dγ if 0 ≤ u ≤ 1
+∞ otherwise.
TheoremThe functionals Jε Γ-converge for the weak L2γ(X ) topology to
cWF where cW is the usual constant.
![Page 45: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/45.jpg)
Some observations
The functional F appears in an alternative proof of theisoperimetric inequality by functional inequalities :
U
(∫
X
u dγ
)≤ F (u).
![Page 46: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/46.jpg)
Some observations
The functional F appears in an alternative proof of theisoperimetric inequality by functional inequalities :
U
(∫
X
u dγ
)≤ F (u).
If u := TtχE , letting t → 0 we get the isoperimetric inequality.
![Page 47: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/47.jpg)
Some observations
The functional F appears in an alternative proof of theisoperimetric inequality by functional inequalities :
U
(∫
X
u dγ
)≤ F (u).
If u := TtχE , letting t → 0 we get the isoperimetric inequality.
⇒ the theory of Bakry-Ledoux on functional inequalites andconvergence to equilibrium in diffusion process.
![Page 48: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/48.jpg)
Convexity of minimizers of
variational problems in Wiener
spaces
![Page 49: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/49.jpg)
Convexity of minimizers of variational problems in Wienerspaces
Problem : Let g ∈ L2γ(X ) be a convex function and v ∈ [0, 1], isthe minimizer of
minγ(E)=v
Pγ(E ) +
∫
E
g dγ
convex ?
![Page 50: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/50.jpg)
Convexity of minimizers of variational problems in Wienerspaces
Problem : Let g ∈ L2γ(X ) be a convex function and v ∈ [0, 1], isthe minimizer of
minγ(E)=v
Pγ(E ) +
∫
E
g dγ
convex ?
By the co-area formula, for ”large” v , this minimizer is a level-setof the minimizer of
∫
X
|Dγu|+1
2
∫
X
(u − g)2 dγ.
![Page 51: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/51.jpg)
Problem : Let F : H → R ∪ +∞ and g ∈ L2γ(X ) be two convexfunctions we want to study the convexity of the solution of
minu∈L2γ(X )
J(u) :=
∫
X
F (Dγu) +1
2
∫
X
(u − g)2 dγ.
![Page 52: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/52.jpg)
Strategy of proof
Approximate the infinite dimensional problem by finitedimensional ones.
Prove the convexity in the finite dimensional case.
Pass to the limit.
![Page 53: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/53.jpg)
The finite dimensional problem
TheoremF : Rm → R ∪ +∞ and g ∈ L2γ(R
m) convex then the solution of
minu∈L2γ(R
m)
∫
Rm
F (Dγu) +1
2
∫
Rm
(u − g)2dγ
is convex.
Idea of proof : construct convex sub- and super-solutions and use aresult of Alvarez, Lasry and Lions to construct a solution.
![Page 54: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/54.jpg)
Relaxation and Representation formulas
DefinitionFor u ∈ L2γ(X ),
∫
X
F (Dγu) := supΦ∈FC1
b(X ,H)
∫
X
−u divγ Φ− F ∗(Φ) dγ.
TheoremFor u ∈ BVγ(X )
∫
X
F (Dγu) =
∫
X
F (∇u)dγ +
∫
X
F∞
(dDs
γu
d |Dsγu|
)d |Ds
γu|
![Page 55: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/55.jpg)
Proposition
Let F be a proper lsc convex function then the functional∫XF (Dγu) is the relaxation of the functional defined as∫
XF (∇u)dγ for u ∈ W
1,1γ (X ).
If F has p-growth, then it is also the relaxation of the functional∫XF (∇u)dγ defined on the smaller class FC1
b(X ).
![Page 56: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/56.jpg)
The infinite dimensional problem
Idea of proof : Let gm := Emg and um be the minimizer of
minu=Emu
Jm(u) :=
∫
X
F (Dγu) +1
2
∫
X
(u − gm)2dγ
by the finite dimensional Theorem, um is convex and um u ⇒ u
is convex.
![Page 57: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/57.jpg)
The infinite dimensional problem
Idea of proof : Let gm := Emg and um be the minimizer of
minu=Emu
Jm(u) :=
∫
X
F (Dγu) +1
2
∫
X
(u − gm)2dγ
by the finite dimensional Theorem, um is convex and um u ⇒ u
is convex.∀u ∈ FC1
b(X ),
J(u) = limm→+∞
Jm(u) ≥ limm→+∞
Jm(um) ≥ J(u)
![Page 58: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/58.jpg)
The infinite dimensional problem
Idea of proof : Let gm := Emg and um be the minimizer of
minu=Emu
Jm(u) :=
∫
X
F (Dγu) +1
2
∫
X
(u − gm)2dγ
by the finite dimensional Theorem, um is convex and um u ⇒ u
is convex.∀u ∈ FC1
b(X ),
J(u) = limm→+∞
Jm(u) ≥ limm→+∞
Jm(um) ≥ J(u)
⇒ By the relaxation Theorem, u is the minimizer of J.
![Page 59: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/59.jpg)
For F one homogeneous with linear growth, we can define theanisotropic perimeter
PF (E ) :=
∫
X
F (DγχE )
then
TheoremFor v large enough, the minimizer of
minγ(E)=v
PF (E ) +
∫
E
g dγ
is unique and convex.
![Page 60: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/60.jpg)
Some perspectives
Look for more accurate algorithms solving the Primal-Dualsystem.
Investigate the case of existence of compact solutions toκ = g when the mean of g is positive.
Look for analog problems in quasi-periodic/stochastic media.
Better understand the link between symmetrizations andfunctional inequalities.
Study representation formulas for more complex integrals.
Define a mean curvature flow in the Wiener space.
...
![Page 61: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/61.jpg)
”Il calcolo delle variazioni e [...] una foresta da esplorare, piuttostoche un palazzo da costruire.”
Ennio De Giorgi
”La tour de Babel” P. Bruegel
![Page 62: Quelques applications des fonctions `a variation born´ee](https://reader031.vdocument.in/reader031/viewer/2022012103/616a113a11a7b741a34e711b/html5/thumbnails/62.jpg)
”Les bulles de savon” J.B.S. Chardin
Merci pour votre attention !